Table 3

Multivariate-adjusted HRs and 95% CIs for the association between fatty pancreas and T2DM incidence, evaluated by P − S, P/S ratio, and quartile of pancreatic attenuation

All of the participants (n = 813)
Model 1Model 2Model 3
HR (95% CI)P valueHR (95% CI)P valueHR (95% CI)P value
P − S (continuous)1.00 (0.98–1.02)0.93
P/S (continuous)1.07 (0.29–3.92)0.92
Reference group1.00 (reference)
Mild group0.92 (0.39–2.18)0.85
Intermediate group1.18 (0.54–2.60)0.68
Severe group1.03 (0.47–2.23)0.95
Age (continuous)1.04 (1.01–1.07)0.0181.04 (1.01–1.07)0.0181.04 (1.01–1.07)0.022
Men1.77 (0.69–4.52)0.231.77 (0.69–4.52)0.231.75 (0.69–4.49)0.24
BMI (continuous)1.18 (1.09–1.29)<0.0011.18 (1.09–1.29)<0.0011.18 (1.08–1.28)<0.001
Liver attenuation (continuous)0.95 (0.93–0.97)<0.0010.95 (0.93–0.97)<0.0010.95 (0.93–0.97)<0.001
Alcohol intake ≥20 g/day1.43 (0.82–2.48)0.201.43 (0.82–2.48)0.211.42 (0.82–2.47)0.21
  • Model 1, model 2, and model 3 were based on the P − S, P/S ratio, and quartile pancreatic attenuation, respectively. Cox proportional hazards models were used to estimate the HRs, 95% CIs, and P values. Spleen attenuation was used as a control in model 1 and model 2. In model 3, all of the participants were divided into the reference group, mild group, intermediate group, and severe group on the basis of quartile pancreas attenuation on unenhanced CT scanning in descending order. The participants in the severe group, who had the lowest pancreas attenuation seen on CT scanning, were considered to have the largest amount of fat in the pancreas, whereas the participants in the reference group, who had the highest pancreas attenuation seen on CT scanning, were considered to have the lowest amount of fat in the pancreas.